Talking photo script templates: 12 short formats

These 12 original templates are starting points for a single portrait speaking a short message. They are not customer testimonials, measured performance examples, or recordings. Use them for welcomes, teaching, invitations, personal greetings, support instructions, and team updates. Replace the bracketed fields with approved details before recording or creating speech.

Choose the message before choosing the voice

Write the audience and the next useful action at the top of your draft. A new learner needs to know where to begin; an event attendee needs a date and location; a support customer needs the next approved step. A portrait cannot demonstrate an interface change or prove a product claim by speaking about it. Put essential evidence beside the presenter, or choose a screen recording when seeing the action is the whole point.

For demonstrations, use the product video script templates. This library focuses on short spoken messages; the photo-to-talking-video walkthrough covers the image-and-audio production steps.

Use a portrait and voice you are authorized to use. If the speaker is synthetic, do not write a real person’s endorsement, experience, or account decision that they did not approve. For support and internal updates, keep private details out of clips that might be forwarded. Choose a neutral presenter when the message does not need a specific person’s likeness.

How to adapt each template

The quoted text is the spoken script. The note underneath is for the writer or producer and should stay outside the speech input. Brackets are replacement instructions, not voice-engine commands. Replace [Name] with the final name, expand unfamiliar acronyms, and write dates in the form you want heard. For example, write “October ninth” instead of expecting a voice to interpret “10/9” correctly for your audience.

Read the filled version aloud once without rushing. If you stumble, rewrite the phrase before choosing a different voice. Keep URLs, reference numbers, and detailed instructions on screen or in an accompanying message when reading them aloud makes the clip harder to follow. Every script below has one main action; remove extra requests rather than squeezing them into the closing sentence.

12 reusable talking-photo scripts

1. A first-step welcome

For a new member who needs a place to start.

“Welcome to [Community], [Name]. Start with [first step]. You do not need to finish everything today. If you get stuck, use [help channel]. We are glad you are here, and we look forward to seeing what you make.”

Adaptation note: Replace the first step with one visible action, not a whole onboarding checklist. Confirm the help channel exists and is appropriate for this audience.

2. A course opening

For a learner starting a specific lesson.

“In this lesson, you will learn how to [skill]. We will use [example] to make the idea concrete. Have [material] ready before you begin. When you finish, try [practice task] and check your answer against [reference].”

Adaptation note: The promised skill must match the lesson. If no answer key exists, replace the last instruction with an available review method.

3. A lesson recap

For someone who has already seen the explanation.

“The main idea is [takeaway]. In our example, [specific observation] showed why it matters. When you face [situation], start by [first decision]. Before you move on, explain that choice in your own words.”

Adaptation note: Use an observation from the actual lesson. A recap should reinforce one concept, not introduce several new definitions after the lesson ends.

4. A practice prompt

For an exercise that asks the viewer to act.

“Try this: [small task]. First, notice [thing to inspect]. Then choose [decision]. Pause here while you work. When you are ready, compare your choice with [worked example] and identify one thing you would change.”

Adaptation note: “Pause here” is a spoken instruction to the viewer. Keep the exercise and answer visible or available elsewhere; do not expect the portrait to display them.

5. An event reminder

For an attendee who needs the essential logistics.

“[Event] starts on [spoken date] at [time and time zone]. Join us at [place]. Bring [item], and arrive by [arrival time] if you need [reason]. The full details are in [message location]. See you there.”

Adaptation note: Check date, time zone, access details, and required items against the event notice. Put long joining links in the accompanying message instead of reading them.

6. A podcast or interview teaser

For a listener deciding whether to hear the full conversation.

“What changes when [specific question]? In this conversation, [Guest] shares [topic they actually discuss]. We also look at [second topic]. Listen to the full episode at [destination], and bring your own answer to that opening question.”

Adaptation note: Use topics present in the approved episode. Avoid invented quotations, sensational claims, or promises that the conversation will solve a problem it only explores.

7. A personal birthday greeting

For one recipient, with a specific shared detail.

“Happy birthday, [Name]. I still smile when I remember [shared moment]. I hope this year gives you more time for [interest] and the people you love. Enjoy your day. I am cheering you on.”

Adaptation note: Include a detail the sender can honestly use and the recipient is comfortable seeing. Personalization should not expose a private memory to a public audience.

8. A thank-you message

For appreciation tied to a real contribution.

“Thank you for [specific contribution], [Name]. It helped us [verified effect]. I especially appreciated [observable detail]. We could not have done this part the same way without your help. Thank you for making time for us.”

Adaptation note: Replace the effect with something known, such as finishing a workshop setup. Do not turn appreciation into an unverified revenue or impact claim.

9. A support next-step message

For an approved instruction, without private account details.

“Hi, [Name]. The next step is [approved action]. You can find it in [location]. If [specific condition] happens, reply through [support channel] with [safe information]. We will review that information before deciding what to do next.”

Adaptation note: Confirm the action and channel with the support owner. Do not imply a refund, repair, deadline, or resolved issue unless that decision is already confirmed.

10. A short FAQ answer

For a repeated question with a bounded answer.

“You asked whether [question]. The answer is [verified answer]. This applies when [condition]. If your situation is different, check [reference] or contact [owner]. Start with [next step] so you have the right information.”

Adaptation note: Name the condition that changes the answer. Keep exceptions that affect the viewer’s decision; remove irrelevant background before shortening the actual answer.

11. A community invitation

For participation with a clear purpose and low-pressure action.

“We are gathering [audience] to talk about [topic]. Bring one question or example you want to share. You are welcome to listen first. If you would like to join, use [signup location] by [date].”

Adaptation note: Check participation rules and signup details. Do not promise confidentiality, eligibility, or access that the organizer has not agreed to provide.

12. A team handoff update

For a teammate who needs the state and next owner.

“[Task] is now at [confirmed stage]. [Owner] has the next action: [action]. The current blocker is [blocker or “none”]. Find the files and notes in [location]. Please confirm [specific decision] before [next milestone].”

Adaptation note: Keep “drafted,” “approved,” and “released” distinct. Replace the blocker placeholder with a spoken phrase and confirm the owner before circulating the clip.

Plan duration with arithmetic, then measure the audio

A word budget is a planning estimate, not a guarantee of clip length. Estimated speech seconds = word count × 60 ÷ assumed words per minute. At an assumed 120 words per minute, 40 words take 20 seconds before extra pauses. At an assumed 150 words per minute, the same 40 words take 16 seconds. These are arithmetic scenarios, not measured Magic Hour voice speeds.

For a 30-second slot, those assumptions produce 60 or 75 words before allowance for pauses. Use fewer words when names, numbers, emphasis, or a quiet ending need more space. Count the filled version: [time and time zone] may become several spoken words. Time the actual recording or generated speech, including intentional silence, before selecting the audio segment for the video.

A worked rewrite: clear logistics without rushing

Illustrative first draft: “We are excited to announce that our upcoming community event will take place next Friday evening, and we would love for you to attend and bring a question while also remembering to sign up and check the joining information before the event begins.”

Illustrative revision: “Our community workshop is Friday at six p.m. Pacific time. Bring one question you want to discuss. Sign up in the event message; you will find the joining details there.”

The revision puts the event and time first, separates preparation from signup, and moves access details out of the spoken line. “Friday” and the time are fictional example details, not an event announcement. In a real message, use a date that stays unambiguous when the clip is replayed. Read the revision aloud and verify its duration; no recording or generated result is reported here.

Review the words before creating the talking photo

Make one editing pass for meaning. Does the first sentence answer the viewer’s question? Is the action possible with the information provided? Replace “click here” when the clip may travel without a clickable link. Cut setup phrases such as “we are pleased to inform you” when they delay the actual message.

Make a second pass for speech. Say each name, number, acronym, and date aloud. Split a sentence when you need to backtrack to understand it. Keep pronunciation guidance in a separate note for a human recording; for generated speech, use only pronunciation or pause controls supported by the speech tool you actually choose. Do not assume bracketed stage directions will become silence rather than spoken words.

Finally, listen to the finished audio without looking at the script. Check whether the facts, intended emphasis, pronunciation, and ending survived. If speech sounds rushed, shorten or rerecord it before animating the portrait. If the audio is understandable but the rendered face has a problem, treat that as a separate video review issue. Rewriting correct dialogue is not a guaranteed repair for facial animation.

Use the script in the correct Talking Photo input

Open Magic Hour Talking Photo with your approved portrait. In Create, choose Add Audio, then Text to Speech for the final spoken script or Upload from device for an approved recording. Listen to the resulting audio before generating the video, and confirm that the selected audio segment contains the complete message. Keep stage directions out of the spoken track. The link opens the tool; it does not transfer a script or portrait from this page.

For an API workflow, create or record the audio first. POST /v1/ai-talking-photo requires assets.image_file_path and assets.audio_file_path; start_seconds and end_seconds select the audio segment. The optional style.prompt guides scene or motion in prompted mode. It is not the dialogue input. Do not paste one of these scripts into that field expecting the portrait to say it.

When the clip is ready, review the full video at normal speed and check the downloaded file against the intended message. The AI video quality assurance checklist covers export review. Keep the approved script and matching audio together so a teammate can revise the message without guessing which version was used.

Official Talking Photo references

The product input distinction above was checked against the current official product page, tool documentation, and API schema on October 18, 2026. Script wording and timing scenarios are editorial examples. The source links below support the image-and-audio handoff; they do not validate the examples as performed videos.

https://docs.magichour.ai/tools/video/talking-photo

https://docs.magichour.ai/api-reference/video-projects/ai-talking-photo

To choose a portrait workflow, compare the best talking-photo tools.